Skip to content

Feature request: Native /voice push-to-talk voice mode in Copilot CLI #3635

@MrRishabhJain

Description

@MrRishabhJain

Feature request: Native /voice push-to-talk voice mode in Copilot CLI

Summary

Add a built-in voice input mode to the Copilot CLI, toggled with a /voice slash command, that lets users dictate prompts hands-free (push-to-talk) directly in the terminal. Speech would be transcribed into the prompt input.

Motivation

There is currently no native voice/speech-to-text capability in Copilot CLI. Today the only option is OS-level dictation (e.g. Win+H on Windows) piped/typed into the prompt, which is clunky and disconnected from the CLI workflow.

Both major points of comparison already ship a voice experience:

  • Claude Code added a built-in Voice Mode (/voice to toggle) in ~March 2026. It uses push-to-talk: hold the space bar to speak, release to send; transcription is streamed into the prompt input and is free / does not count against rate limits. (Referenced in the official Claude Code changelog v2.1.160: "Fixed voice mode failing to connect when the project directory or branch name contains non-ASCII or special characters.")
  • GitHub Copilot in VS Code already supports voice via the official VS Code Speech extension (ms-vscode.vscode-speech) — hold-to-speak microphone, a "Hey Code" voice trigger, and optional TTS read-back.

The CLI is the notable gap: Copilot has voice in the editor but not in the terminal.

Proposed behavior

  • /voice toggles voice mode on/off.
  • Push-to-talk (e.g. hold a key to record, release to transcribe and insert into the prompt input).
  • Transcribed text lands in the editable prompt input so the user can review/correct before sending.
  • Clear on-screen indicator when recording / transcribing.
  • Sensible privacy posture (not always-listening; explicit push-to-talk).

Prior art

  • Claude Code Voice Mode (/voice, push-to-talk via space bar).
  • VS Code Speech extension for Copilot Chat (hold-to-speak, "Hey Code").

Notes

No existing issue requesting this was found when searching the repo for voice, speech, microphone, dictation, and audio.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions